Correlations Notes Output Created 12-FEB-2024 23:01:50 Comments Input Data C:\Users\Administrator\Downloads\NCS-data set for assignments (8).sav Active Dataset DataSet1 Syntax CORRELATIONS /VARIABLES=AGE ABUSE /PRINT=TWOTAIL NOSIG FULL /MISSING=PAIRWISE. Resources Processor Time 00:00:00.05 Elapsed Time 00:00:00.05

Correlations Age Frequency of childhood abuse experiences Age Pearson Correlation 1 -.086** Sig. (2-tailed) <.001 N 9282 9153 Frequency of childhood abuse experiences Pearson Correlation -.086** 1 Sig. (2-tailed) <.001 N 9153 9153 ** Correlation is significant at the 0.01 level (2-tailed).

GGraph Notes Output Created 12-FEB-2024 23:24:35 Comments Input Data C:\Users\Administrator\Downloads\NCS-data set for assignments (8).sav Active Dataset DataSet1 Filter Weight Split File N of Rows in Working Data File 9282 Syntax GGRAPH /GRAPHDATASET NAME="graphdataset" VARIABLES=ABUSE AGE MISSING=LISTWISE REPORTMISSING=NO /GRAPHSPEC SOURCE=INLINE /FITLINE TOTAL=NO SUBGROUP=NO. BEGIN GPL SOURCE: s=userSource(id("graphdataset")) DATA: ABUSE=col(source(s), name("ABUSE")) DATA: AGE=col(source(s), name("AGE")) GUIDE: axis(dim(1), label("Frequency of childhood abuse experiences")) GUIDE: axis(dim(2), label("Age")) GUIDE: text.title(label("Scatter Plot of Age by Frequency of childhood abuse experiences")) ELEMENT: point(position(ABUSE*AGE)) END GPL. Resources Processor Time 00:00:07.14 Elapsed Time 00:00:42.18

1. Compute the Pearson correlation, r, and a scatterplot of the two variables you used in the previous assignment. Upload the SPSS output file.

See pages 91-93 of the text for instructions on using SPSS to compute the statistics.

Assume you computed an r = +.50 for the variables, exercise and lifespan.

2. What is the magnitude of the correlation (high, medium, low)?

High, or 'strong' (see page 87, Table 5.3)

3. What is the direction of the correlation (positive or negative)?

Positive

Note: the r value for a negative correlation would have a negative sign in front of it, e.g., -.50.

4. What prediction can you make based on the variables if theircorrelation, r = .50?

People who exercise a lot will live longer.

5. What prediction can you make based on the variables if theircorrelation, r = - .50?

People who exercise a lot will live shorter lives.

6. Describe the scatterplot of the correlation that you computed. Do the dots line up diagonally from left to right, or form a random pattern?

In a scatterplot for r = .50, the points would form a line from the lower left to the upper right, but many points would be scattered around the line. If the correlation was low, r < .30, you would not be able to see a line in a scatterplot, only a random scatter of points.

7. Why can you not conclude that one variable causes the other based on r, even if r=1?

Even if the correlation of lifespan and exercise was perfect, r=1, you cannot infer that exercise causes an increase in lifespan because a 'confounding' variable, like genetics, might cause people to both live longer and exercise more, creating the correlation of exercise and lifespan.

8. What confounding variable in the NCS data set might cause the correlation that you found?

The age of people's parents, if available might be a good measure of a genetic predisposition to longevity.

9. Compute the partial correlation of the two variables controlling for the confounding variable (attach the SPSS output file).

See pages 96-98 of the text for instructions.

10. Given the partial correlation, is the correlation confounded by the control variable? Explain why or why not.

If you computed a partial correlation between lifespan and exercise, controlling for genetics, it would subtract the correlations of genetics with both exercise and lifespan from their correlation with each other. If the partial r is lower than their original r, then genetics confounded their correlation; the correlation between exercise and lifespan is lower after the effects of genetics have been removed.